Cliptate v1.0 · First open-source, free direct-download release

Features and improvements

• Local speech: Transcribe audio/video with Parakeet, Qwen3 ASR, or FireRed, using custom vocabulary, alignment, and speaker attribution. Dictation supports streaming recognition and shortcuts.
• Editing and captions: Trim pauses, review speaker-labelled transcripts, correct and translate captions, and edit and export timelines. Speech synthesis offers preset voices, cloning, emotion control, and voice design.
• Model management: Centralized selection, downloads, preload, cache status, and retries reuse compatible Hugging Face snapshots. Speech models run locally; optional online AI correction or translation uses separately configured services.
• Source and distribution: The complete app is MIT-licensed. Direct GitHub builds need no subscription or purchase and provide Sparkle in-app updates. App Store subscription policy remains unchanged.

Upgrade notes

• Requires Apple Silicon and macOS 15+; Intel Macs are unsupported. Download Cliptate.dmg and verify SHA-256, close the old app, and drag it into Applications. Download model weights separately.
• The v1.0 direct build is still sandboxed and reuses compatible caches; the later sandbox-data migration is not included. App identity stays unchanged: retain existing data and caches. Automatic update checking does not enable automatic installation.

Known limitations

v1.0 does not include the native CLI. App Store builds retain subscription and Accessibility limits. These notes were reconstructed from the historical tag without a fresh test run; the document alone does not establish signing, notarization, or end-to-end update status.